Literature Survey in blind Channel Estimation for MIMO-OFDM

Pranisha P M , Packia Das Y R2

Abstract: Our chief goal in this paper is to produce an
analysis on blind channel estimation method based on the various approach. We address the problem of estimating a distorted signal in the presence of noise in the case that very little prior knowledge is available regarding the nature of the signal, the distortion, or the noise. This paper describes a survey on various techniques for blind channel estimation and issues related to individual operations. The proposed channel estimation approaches employs computationally simple crosscorrelation operations and yields the channel up to a diagonal ambiguity. It does not require channel length information and is not sensitive to additive stationary noise. We outline basic ideas behind several developments, the assumptions and identifiably conditions required by these approaches, and the algorithm characteristics. We refer to the problem as blind estimation without priors, where the term blind captures the notion that signal estimates are obtained blindly with regard to knowledge of the distortion and interference. The proposed paper describes comparison between approaches and methods to estimate the a novel blind channel estimation algorithm for a multiple input multiple output (MIMO).

Designs of receivers that remove channel distortions require either the knowledge of the channel or the access to the training signals. The latter is the choice in many communication systems design. The transmission of training signals decreases communications throughput although, for time invariant channels, the loss is insignificant because only one training sequence is necessary. For time varying channels, however, the loss of throughput becomes a problem [4]. The blind channel estimation means that the channel is estimated without any training sequence; instead, the identification is achieved by using only the channel output with certain a priori statistical information on the input. Such methods can increase the transmission capability due to the elimination of training signals Estimation of a distorted signal in noise is a classic problem that finds important application in areas including, but not limited to, Data communication Radar signal processing Sensor array processing Geophysical exploration Speech processing Image analysis Biomedicine Control systems OFDM systems have recently emerged as good candidates for future generation high-rate wireless systems due to their efficient utilization of bandwidth and robustness to multipath. In this paper we focus on blind channel estimation for OFDM systems. The literature along these lines can be categorized as follows. Single-user single antenna OFDM systems-Each symbol is sent over a different carrier, over which it encounters flat fading. Considering blocks of received and transmitted symbols, one can formulate a MIMO problem, where the channel matrix is diagonal. There isa cornucopia of available methods for blind channel estimation in this case [5][6],[7]. A weakness of single antenna OFDM systems is the lack of multipath diversity. Precoding can spread each symbol over multiple carriers thus increasing diversity. Unitary precoding has been shown to provide maximum spectral efficiency, however, it does not allow for blind channel estimation. A non-unitary linear precoding scheme

Keywords Annalysis, Blind, Estimation, distorted, interference,

MIMO

I. INTRODUCTION

ecently introduced Multiple Input Multiple Output (MIMO) signal transmission schemes are attractive for newly emerging high-speed data transmission wireless communication systems because they offer an increased data throughput (capacity) without increasing operational bandwidth [1], [2]. Their other advantage is they are capable to enhance the quality of signal transmission through the use of transmitter and receiver diversity. These benefits are possible under the condition that the MIMO channel state information (CSI) is available at the receiver. Traditionally CSI is acquired by sending training sequences (pilot signals) evenly spaced along a block of transmit symbols. In order to save the bandwidth and increase spectral efficiency, blind and semi-blind channel estimation methods are applied to obtain the CSI. Inter symbol interference (ISI) is a limiting factor in many communication systems. ISI can arise from timevarying multi-path fading, which can be severe in, for example, a mobile communication system. Other channel impairments that contribute to ISI include symbol clock jitter, carrier phase jitter, etc. To achieve high-speed reliable communication, channel estimation and equalization are necessary to overcome the effects of ISI [3].

was proposed in [8], which trades-off performance for blind channel estimation. Single-user OFDM systems with multiple transmit/receive antennae-Multiple transmit/receive antennas in combination with coding are used to improve diversity and rate. Some representative schemes include the space time block codes (STBC) [9], spacetime trellis codes, and layered space time. In this case, the channel matrix is nondiagonal. A channel estimation scheme for such MIMO systems is the two-input one-output MIMO OFDM system studied in [9], where STBC was applied at the inputs and the system was estimated by exploiting the structure of the codes. This approach relies on transmission redundancy, i.e., each information symbol is transmitted twice in two consecutive time intervals through two different antennas. In [4], the same setup as in [9] was used, and the channel was estimated based on a subspace approach. Multi-user OFDM systems without multi-user interferenceEach user is allocated a disjoint set of subcarriers [3]. Such systems are also referred to as orthogonal frequency division multiple accesses (OFDMA) and have been adopted in IEEE 802.16 standard. In such systems there is no multi-user interference and the channel matrix is diagonal. Thus, any single-user frequency domain channel estimation method can be applied to this scenario. Multi-user OFDM systems with multi-user interference-All users use all available subcarriers inde-pendently. Each user can employ one or more antenna. Inevitably, in this case there is multi-user. Interference and the channel matrix is non-diagonal. The majority of the single-user OFDM channel estimation methods do not apply to this case. There exists only a few channel estimation schemes for such systems. In [2], an OFDM system with cyclic prefix has been considered. Each block was multiplied with a scalar that varied periodically between blocks. Cyclic statistics of the received blocks, before the removal of the cyclic prefix, were used to yield the channel estimate. This kind of precoding results in variable power between blocks. A semi blind channel estimation method was proposed for multi-user OFDM systems that employ zero-padding instead of cyclic prefix. We should note here that the channel estimation methods developed for multiuser multi-antenna systems can be applied to most single user multi-antenna systems. By reviewing recent surveys [5], [10], the purpose of this paper is to review some blind channel estimation approaches. We provide a systematic summary of some algorithms in the area of blind channel estimation. Various existing algorithms

are classified into the moment-based and the maximum likelihood (ML) methods. If input is assumed to be random with prescribed statistics, the corresponding blind channel estimation schemes are considered to be statistical. On the other hand, if the source does not have a statistical description, or although the source is random but the statistical properties of the source are not used, the corresponding estimation algorithms are deterministic [10]. The paper further organized as follows: Section II proposes Basic Description. Section III Algorithms Description while statistical result is shown in section IV and conclusion is defined in section V.
II. BASIC DESCRIPTION

A mathematical model describing this problem is r = H(s) +w 1

Here the observed vectorris modeled as a signal vector sdistorted by function H()Different estimation problems can be characterized by the assumptions placed on H(), s, and w. Though central limit theorem arguments are commonly used to justify a Gaussian noise model for w, the assumptions on H() and scan differ significantly from one application to the next. For example, is H() completely known? If not, is it because H() is inherently random? Assuming random H(), do we know its distribution? If not, can variabilitys in the distribution or structure of H() be described with a small set of parameters? Similar questions can be asked about the signals or about the noise w when the Gaussian assumption is not adequate. In general, the introduction of accurate prior assumptions about H(),s, and w, allows the design of estimators with increased performance (though perhaps more complicated implementation). If prior assumptions are inaccurate, however, estimator performance can suffer significantly. There exist many applications of estimation theory where prior knowledge is lacking and accurate assumptions are hard to come by. In this dissertation we focus on a relatively extreme lack of knowledge. Specifically, we assume that 1. H() is linear but otherwise unknown,

Blind Channel Estimation

Statistical Method Maximum Likelihood Moment Method

Deterministic Method Maximum Likelihood Moment Method

Subspace Methods

Moment Matching
Fig 1: Classification of blind channel estimator [7]

2. the signal vectors is composed of statistically independent and identically distributed random variables with unknown distribution, and 3. the noise vector w is composed of identically distributed random variables, statistically independent ofs, with unknown distribution. One of key distinguishing features of our problem setup is the unknown linear structure of H(). In the five previously mentioned applications, this feature often corresponds to a lack of knowledge about the dispersion pattern of the communication channel, geometry of the array, shape of the seismic wavelet, blurring characteristics of the lens, or speaker mixing matrix, respectively. The term blind estimation has been used to describe problems of this type since the estimation of the signal is performed blindly with respect to the channel and noise characteristics. For similar reasons, blind identification denotes the identification of the unknown system H() in the context of unknown signal and noise. Though the term blind estimation is sometimes used to describe situations where, for instance, H() is unknown but the signal distribution is known, we stress that our problem setup is different in that it assumes no prior knowledge about distribution of signals and structure of distortion H() (with the exception, respectively, of independence and linearity).

II. REVIEW ON VARIOUS SYSTEMS

Many recent blind channel estimation techniques exploit subspace structures of observation. The key idea is that the channel vector is in a one- dimensional subspace of the observation statistics. These methods, which are often referred to as subspace algorithms, have the attractive property that the channel estimates can often be obtained in a closed form from Optimizing a quadratic cost function. Subspace methods can sometimes be considered part of the moment methods. Several blind channel estimation methods have been described in [3] [4]. These methods are based on the subspace algorithm [5], which utilizes the orthogonally between the channel matrix and the Sylvester matrix-formed noise subspace. There are several drawbacks of subspace-based MIMO channel estimation methods. Firstly, they suffer from so-called multidimensional ambiguity. Several training symbols are necessary to eliminate this ambiguity; Secondly, in order to compensate for extra degrees of freedom in the noise subspace, when the number of transmit antennas are smaller than the number of antennas at the receiver, the methods presented in [6] [7] require pre-coding. Thirdly, eigenvalue decomposition is needed to recover the channel matrix without ambiguity. This leads to high implementation complexities.

In [11] [12], a semi-blind channel estimation approach employing orthogonal pilot maximum likelihood (OPML) estimator has been proposed. The approach performs SVD onto the received signal correlation matrix estimating the whitening matrix of channel. Using the information carried by the whitening matrix, the OPML estimator shows a 1dB improvement of bit error rate (BER) compared to the conventional least squares (LS) training scheme if the same length of training sequence is used. However, it still requires training symbols to achieve the same performance as LS. Furthermore, SVD has to be applied twice to obtain the whitening and rotation matrices. This leads to the increased computational complexity. The authors of [10] have presented a new singular value Decomposition (SVD) blind channel estimation scheme employing a simple block pre-coding structure. The advantage of the presented approach is that the CSI can be recovered without ambiguity if the proper modulation is applied. Another gain of this scheme is that no block decoder is needed at the receiver. These advantages are accomplished at the expense of the coding rate, which decreases with the number of transmitting antennas. Therefore, the scheme proposed in [8] is suitable only for a 2x2 MIMO system to avoid a waste of precious bandwidth. There are also several blind methods for identifying frequency-selective MIMO channels (for example, see [13] and references therein) which do not assume any space-time coding and, therefore, are not able to take advantage of the orthogonal structure of the codes used. There are several promising approaches to channel estimation in space-time coded MIMOOFDM systems [13]-[14]. However, the techniques [14] are only applicable to the case of two transmit and one receive antennas, while the approach of [9] generally requires the number of receive antennas to be not less than the number of transmit antennas 1. Obviously, the latter restriction may be critical for the downlink mode. The semi blind approach of [10] re-quires to transmit pilot symbols for a part of subcarriers used. Another common approach to channel estimation in MIMO-OFDM systems is to estimate flat fading channels at each sub-carrier independently in the frequency domain [15]. However, this method does not enable coherent processing across the subcarriers and may suffer from a high computational cost in the case when the total number of subcarriers is large.

Fig 2:MIMO-OFDM system model with Mt transmit and Mr receive antennas. Linear block precoding (LBP) has been used extensively for increasing throughput and diversity gains in MIMO OFDM systems. The design of codes to achieve these goals has recently become an area of huge interest and vast potential. For instance, algebraic number theory [16], coding theory [15], and iterative greedy optimization [17] have being used to design optimal codes. However, very little work has been done on developing LBP schemes that allow channel estimation.

A. Subspace-based Identification
A blind identification procedure consists in estimating the L(M+1)1 vector H of channel coefficients: H = [H(0)T,,H(L-1)T]T 1 From (5), we can see that: Yn = HnXn+Wn 2

So, the identification is based on the LNLN autocorrelation matrix of the measurement vector: Ryy = E{YnYnH} 3

Since the additive noise is assumed independent of the emitted sequence, the autocorrelation matrix is expressed: Ryy = HnRkHnH+Rww 4

The source covariance matrix Rk has dimension (N +M)(N+M), and is assumed to be full rank but otherwise unknown. The noise covariance matrix Rww is of size LNLN. To ease the derivations, the noise is assumed to be white.

B. Subspace Decomposition

Let e0 e1 eLN-1 denote the eigenvalues of Ryy Since Rk is full rank, the signal part of autocorrelation matrix HnRkHnH has rank M+N, since Rww = d2I, Hence: ei > d2 for ei = d2 for i =0,,M+N-1 i = M+N,,LN-1 5

The autocorrelation matrix is thus also expressed as:

Ryy = S diag (e0,e1,,eM+N-1)SH+d2GGH S = [S0,S1,..,SM+N-1] LN X (M+N) G = [G0,G1,,GLN-M-N-1] LN X (LN-M-N)

constraint avoiding the trivial solution H =0. Different constraints on H provide different solutions. We have classically considered minimi zation subject to linear and quadratic constraints: Quadratic constraint: Minimize q(H) subject to | H|=1. The solution is the unit-norm eigenvector related to the smallest eigenvalue of matrix Q. Linear constraint: Minimize q(H) subject to cHH =1. Where c is a L (M+1)1 vector. The solution is proportional to Q-1C [18]. The first choice is more natural but involves the computation of an additional eigenvector. The second solution depends on the choice of an arbitrary constraint vector c. The computational cost of the second solution is lower since it amounts to solving a linear system rather than extracting an eigenvector [18].

The columns of matrix S span the signal subspace (dimension M+N), while the columns of G span its orthogonal complement, the noise subspace. Also, the signal subspace is the spanned by the columns of the filtering matrix HN. By orthogonality between the noise and the signal subspace, the columns of HN are orthogonal to any vector in the noise subspace. Hence, we have: GiHHN = 0 0 i<LN-M-N 7

Under some conditions detailed in the theorem below, the noise subspace (matrix G ) uniquely determines the channel coefficients up to a multiplicative constant. Theorem 1: assume that N M and matrix HN-1 is full rank. Let HN be a new nonzero matrix with same dimension as HN .The range of new matrix is included in the range of H N iff the corresponding H and H are proportional. Hence, both of matrices share the same column space iff they are proportional. We show in the next section how by using this theorem the channel coefficients can be esti mated even in the cases where Rd is full rank and unknown.

D. Signal Subspace
It is shown before that minimizing a constrained quadratic form involving the noise eigenvectors give the channel coefficients. This quadratic form is equivalently rewritten in terms of the signal eigenvectors as:

11 E. Deterministic Subspace Approach

C. Subspace-Based Parameter Estimation Scheme

Sample estimates Gi, of the noise eigenvectors are available and (7) is solved in the least squares sense and leads to minimize the following quadratic form:

8
As you can see, q(H) depend on vector H rather than on the filtering matrix H N. This is conveniently done by application of following Lemma, which requires the following notations. Notations : Let V (0) ,,V (L-1) be L arbitrary N1 vectors and let V be the LN1 vector defined as V=[V(0)T,,V(L-1)T]T.

Subspace method based on the property that the channel is in a unique direction. It may not be robust against modeling errors, especially when the channel matrix is close to being singular. The second disadvantage is that they are often more computationally expensive. Without presence of noise, the estimator produces the exact channel using only a finite number of samples if the identifiability condition is satisfied. Therefore, these methods are most effective at high SNR and for small data sample applications. On one hand, deterministic methods can be applied to a much wider range of source signals; on the other hand, not using the source statistics affects its as ymptotic performance

E. LMMSE Channel Estimation

LMMSE is widely used in the OFDM channel estimation since it is optimum in minimizing the MSE of the channel estimates in the presence of AWGN. LMMSE uses additional information like the SNR. LMMSE is a smoother/ interpolator /extrapolator, and hence is very attractive for the channel estimation of OFDM based syst ems with pilot subcarriers. However, the computational complexity of LMMSE is very high due to extra information incorporated in the estimation technique. The LMMSE estimator (19) is very complex since a matrix inversion is needed every time the data in changes. We reduce

By theorem 1, if true autocorrelation matrix was available, the true channel coefficients are the unique (up to a scalar factor) vector H such that q(H) =0. In contrast, when only an estimate of the autocorrelation matrix is available, the quadratic form has not exactly rank L(M+1). Hence, estimation of H can be obtained by minimizing q(H) subject to a properly chosen

the complexity by averaging over the transmitted data. We replace the term (XXH)-1 in with its expectation E{(XXH)-1}. V. CONCLUSION
Channel estimation is a standard linear system identification problem with the training sequence as the pilot input signal. In many applications, the pilot signals may not be easy to use or they may present an extra problem, for example requiring more bandwidth in communication systems. Blind channel estimation and equalization eliminates the need for a pilot signal and simplifies the requirements for channel estimation and equalization. In particular, recent developments in blind estimation research have led to a class of rapidly converging and data efficient algorithms th at can effectively estimate the channel with a small number of data points. In this paper, we reviewed some of the basic approaches in blind estimation.

[14]

T.-H. Chang, W.-K. Ma, and C.-Y. Chi, Semiblind OSTBC-OFDM detection in block fading channels,Proc. ICASSP07 , April 2007, Honolulu, Hawaii, USA, vol. 3, pp. 309-312.

[15] T. Zemen and C. F. Mecklenbrauker, Timevariant channel estimation using discrete prolate spheroidal sequences, IEEE Trans. Signal Process. , vol. 53, pp. 35973607, Sep. 2005. [16] D.C. Popescu and C. Rose, Interference avoidance and multiuser MIMO systems, Int. J. Satell. Commun. Network. vol. 21 pp. 143-161, Oct. 2003. [17] H. El Gamal, A.R. Hammons, Y. Liu, M. P. Fitz, and O. Takeshita, On the design of space time and space frequency codes for MIMO frequency selective fading channels, IEEE Trans. on Information Theory, vol. 49, no. 9, pp. 2277-2292, Sep. 2003. [18] Norooz Motamedi, Blind Channel Estimation using Subspace and LMMSE for SIMO OFDM Systems: A Review, IEEE Trans. on Information Theory, vol. 49, no. 9, pp. 2277-2292, Sep. 2003.

REFRENCES
[1] I. Telatar, Capacity of multiple antenna Gaussian channels, Eur. Trans. Telecomm., Vol. 10, No., pp.585-595, Nov./Dec. 1999. [2] G. J. Foschini and M. J. Gans, On limits of wireless communications in a fading environment when using multiple antennas, Wireless Personal Commun., Vol. 6, pp. 311-335, 1998. [3] Tong L., et al, Blind Identification and Equalization based on Second-Order Statistics: A Time Domain Approach, IEEE Transaction on Information Theory, Vol.40, No.2, pp.340-349, March 1994. [4] Tong L., Perreau S., Multichannel Blind Identification: From Subspace to Maximum Likelihood Methods, Pro ceeding of the IEEE, Vol.86, No. 10, pp.1951-1968, October 1998. [5] Y. Zeng and T. Ng, A semi blind channel estimation method for multiuser multiantenna OFDM systems, IEEE Trans. on Signal Processing, vol. 52, no. 5, pp. 1419-1429, May 2004. [6] S. Zhou and G. B. Giannakis, Finite alphabet based channel estimation for OFDM and related multicarrier systems, IEEE Trans. on Signal Processing, vol. 49, no. 8, pp. 1402-1414, Aug. 2001. [7] B. Muquet, M. de Courville, and P. Duhamel, Subspace based blind and semi blind channel estimation for OFDM systems, IEEE Trans. on Signal Processing, vol. 50, no. 7, pp. 30653073, July 2002. [8] A. P. Petropulu, R. Zhang and R. Lin, Blind OFDM Channel Estimation through Simple Linear Precoding, IEEE Trans.on Wireless Communications,vol. 3, no. 2, pp. 647-655, Mar. 2004. [9] S. M. Alamouti, A simple transmit diversity technique for wireless communications, IEEE J. Select. Areas Commun.,vol. 16, pp. 1451-1458, Oct. 1998. [10] Tong L., Perreau S., Multichannel Blind Identification: From Subspace to Maximum Likelihood Methods, Pro ceeding of the IEEE, Vol.86, No. 10, pp.1951-1968, October 1998. [11] A. Jagannatham and B. D. Rao, A semi-blind technique for MIMO channel matrix estimation, Proc. Of IEEE 4 th Workshop on singal processing in Wireless Communication (SPAWC 2003), #582, Rome, Italy, 2003 [12] A. Jagannatham and B. D. Rao, Constrained ML algorithms for semi-blind MIMO channel estimation, Proc. of IEEE Communication Society Globecom, 2004 [13] Z. Liu, G. B. Giannakis, S. Barbarossa, and A. Scaglione, Transmit-antennae space-time block coding for generalized OFDM in the presence of unknown multipath, IEEE J. Sel. Areas Commun., vol. 19, pp. 1352-1364, July 2001

Packia dhas. Y.R is currently working as Associate Professor in PET Engineering College, Vallioor.He have an experience of 10
years and two months. He is also a member of ISTE.

P of

M Pranisha received the B.E in Electronics and Communication Engineering from VINS Christian College Engineering in 2011 and currently doing M.E. Degree in Communication Systems under Anna University, Tirunelveli. His Area of Interest includes signal Processing and Image processing.